1. Field of the Invention
The present invention generally relates to the measurement of the electrical characteristics of a twisted-pair cable and, more particularly, to a multi-port mechanical system for twisted pair cable applications.
2. Description of the Related Art
When using and designing high frequency multiport networks, such as data cables and connecting hardware, it is often necessary to characterize both the differential performance as well as the common mode performance of the data interface. With the recent increase in operating frequencies, there has arisen a very strong need to understand the high frequency coupling between the various modes of propagation within cables and connecting hardware. Unfortunately, test and measurement hardware suitable for use at low frequencies is unsuitable at the frequencies presently contemplated because previously disregardable errors become significant.
Errors associated with network analyzer measurements can result from either non-ideal conditions in the measurement equipment or from the effects of the mechanical fixtures, such as cables and connectors, which are required to connect a device under test (DUT) to the test equipment. With particular respect to the mechanical fixtures, errors can result from motion of the fixture between measurements, coupling between fixtures, and coupling between conductors from the twisted pair cable in the fanout region (i.e., where the conductors are untwisted so they can be connected to the input ports of the network analyzer). Although these errors are relatively minor at low frequencies, at high frequencies, they corrupt calibration and measurement data to an undesired degree.
Conventionally, the differential performance of a system has been measured by connecting a balun to each port of the system while using a network analyzer to determine the actual performance parameters. Such a balun is disclosed in U.S. Pat. No. 4,717,896 to Graham which discloses a balun for intercoupling a partially unbalanced line with a substantially balanced line having similar line impedances. The balun is capable of passing differential signals over a wide bandwidth while substantially attenuating any common mode signals. The disadvantage of this approach is that the balun itself corrupts the measurement. Bandwidth limitations limit the frequency range of the measurement and imbalances in the balun contaminate differential mode signals used to probe the various ports of the DUT. Additionally, when attempting to determine coupling related to various modes of propagation, balun characteristics need to be subtracted from the measurements. To date, no credible procedure has been developed to accomplish this type of calibration.
In addition to the problems associated with baluns, conventional systems and methods for determining the performance parameters of cables and connecting hardware do not possess the precision necessary to extract all of the coupling parameters. Specifically, mode conversions are not measured accurately. Further, the conventional fixtures used to connect the DUT to the measurement device are unrefined and corrupt the measurements at high frequencies. Compounding these deficiencies is the inadequacy of the calibration procedures used to compensate for imperfections of the fixtures.
Conventional systems also fail to provide mechanical fixtures which minimize pair fan out when connecting a twisted-pair cable to the measurement system. In addition, due to the flexible nature of the mechanical fixtures of the conventional systems, it is difficult to reproduce measurements accurately. Because of these deficiencies, conventional systems are incapable of utilizing calibration equipment which can overcome the deficiencies of the presently available measurement procedures. As a result, conventional systems fail to adequately measure differential and common mode return loss, crosstalk, attenuation and mode conversions between differential and common mode signals in a DUT, such as a twisted-pair cable.
A conventional modal decomposition system is the Hewlett Packard 4380s measurement system. While suitable for some differential measurements, this unit cannot be accurately calibrated for measurements involving common mode parameters. Further, the mechanical fixtures provided with this system to connect the DUT to the analyzer ports are inadequate for high frequency applications and the coupling cables are free to move during the calibration processes. As a result, the noise floor is substantially corrupted and it is difficult to obtain reproducible measurements at high frequencies. Moreover, the symmetries expected in the measurement of reciprocal networks are absent, thus further indicating measurement problems. Finally, this system cannot measure mode conversions in high frequency systems.
It would therefore be advantageous to provide a system for making appropriate measurements of the high frequency characteristics of a DUT, e.g., a twisted cable, which does not suffer from errors present in conventional measurement systems.
The above and other problems are overcome in a multi-port modal decomposition system which provides a rigid interconnection between a DUT and a network analyzer. The multi-port modal decomposition system is a mechanical device which may be used both during calibration measurements and when measuring the high frequency characteristics of various electrical components.
According to the invention, a novel mechanical fixture is used to provide a rigid connection between a multi-conductor cable or other connecting hardware and measurement equipment. The fixture includes a test head consisting of a plurality of unbalanced ports which are contained in a generally circular area whose diameter is close to that of the cable to be tested. As a result, fan out of the conductors at the point where they connect to the ports in the test head, which can corrupt measured data, is minimized. The test head is rigidly connectable to a multiport RF switching network, preferably by the use of a removable xe2x80x9csnapxe2x80x9d connection. Additional ports on the test head can be provided for use during a calibration process.
In one embodiment of the invention, an electrical cap is provided for holding the conductor ends in place. The cap mates with the test head to rigidly connect the conductor to the measurement ports while retaining the xe2x80x9cfanned outxe2x80x9d portion of the cable in a rigid and known configuration. In a further embodiment, a slidable shield is provided which extends from within the test head and serves to electrically isolate the individual pairs of conductors in the fan-out region of the cable from each other. The rigid mechanical connections between the DUT and the measurement equipment advantageously provides the low noise floor and reproducibility needed to achieve precise results.